/ / Little Brother, News

I just finished my podcast reading of my latest story, “Epoch,” which Mark Shuttleworth commissioned for my upcoming short story collection/experiment, With a Little Help. It’s the story of the sysadmin charged with shutting down the first and only functional AI, which no one can figure out a reason to save — and it’s the story of the AI’s bid to save its own life by fixing the Unix 32-bit rollover problem.

The podcast is in eight parts — I started reading it before I’d finished the story, so there’s some minor inconsistencies that’ll be fixed in the final cut. Next up I’ll be reading “Martian Chronicles,” my young adult story about free-market ideologues colonizing Mars, and the video games they play on the way to the Red Planet.

The doomed rogue AI is called BIGMAC and he is my responsibility. Not my responsibility as in “I am the creator of BIGMAC, responsible for his existence on this planet.” That honor belongs to the long-departed Dr Shannon, one of the shining lights of the once great Sun-Oracle Institute for Advanced Studies, and he had been dead for years before I even started here as a lowly sysadmin.

No, BIGMAC is my responsibility as in, “I, Odell Vyphus, am the systems administrator responsible for his care, feeding and eventual euthanizing.” Truth be told, I’d rather be Dr Shannon (except for the being dead part). I may be a lowly grunt, but I’m smart enough to know that being the Man Who Gave The World AI is better than being The Kid Who Killed It.

Not that anyone would care, really. 115 years after Mary Shelley first started humanity’s hands wringing over the possibility that we would create a machine as smart as us but out of our control, Dr Shannon did it, and it turned out to be incredibly, utterly boring. BIGMAC played chess as well as the non-self-aware computers, but he could muster some passable trash-talk while he beat you. BIGMAC could trade banalities all day long with any Turing tester who wanted to waste a day chatting with an AI. BIGMAC could solve some pretty cool vision-system problems that had eluded us for a long time, and he wasn’t a bad UI to a search engine, but the incremental benefit over non-self-aware vision systems and UIs was pretty slender. There just weren’t any killer apps for AI.


Podcast feed